Giffen behavior has been fascinating economists ever since it was first stated in the classical works of Henry Beeke (1800), George Porter (1847), Simon Gray (1815) and Alfred Marshall (1895). In these first interpretations, this behavior was said to be predominantly experienced by consumers whose consumption had been diminished to subsistence levels. Often referred to as “Giffen’s Paradox”, Sir Robert Giffen originally explained it as “a rise in the price of bread makes so large a drain on the resources of the poorer laboring families and raises so much the marginal utility of money to them, that they are forced to curtail their consumption of meat and the more expensive farinaceous foods: and, bread being still the cheapest food which they can get and will take, they consume more, and not less of it” (Marshall A., 1895), implying an upward sloping demand curve and thus defying a fundamental economic principal, the law of demand. However, although the enigma has attracted many theoretical interpretations by, among others, Hicks and Allen (1934), Stigler (1947, 1948, 1950) and Vandermeulen (1972), it was argued that there remained a lack of empirical verification confirming the existence of Giffen behavior (Stigler G., 1947; Prest A., 1948; Dwyer P. and Lindsay C., 1984). Dismissing it as a theoretical feasibility but a practical impossibility (Hicks J., 1956; Stigler G., 1947, 1948), numerous approaches have since been employed to uncover unequivocal evidence of the phenomenon.
One approach to explaining the Giffen paradox, is a purely graphical analysis that uses a combination of indifference curves reflecting a positive own-price effect. This technique can be accompanied by the Slutsky equation and the Hicksian theory of demand which specifies a single constraint household optimization problem where the absolute size of a substitution effect is offset by the absolute size of a larger negative income effect, resulting in Giffen behavior (Hicks J., 1956; Slutsky E., 1956). A second popular method involves specific utility functions that are able to generate Giffen behavior. Examples of these utility functions have been proposed by Wold and Jureen (1953), Spiegel (1994) and Weber (1997). However, Moffatt (2002) claimed that they contradicted the accepted axioms of consumer theory and as a result, discredited them. There also exists a third procedure often used when proving that Giffen behavior only emerges when income levels are low. This approach argues that the consumer’s optimization problem is not only restricted by a budget constraint but also a rationing constraint. Gilley and Karels (1991) were among the first to provide a formal analysis of this idea using a basic two-good example with the assumption that one of the goods has a higher efficiency in contributing to the rationing constraint. Due to the binding nature of both constraints, the more efficient good displays a positive price elasticity and thus the Giffen paradox emerges. As a result of the theory having been derived solely from two binding constraints, Gilley and Karels (1991) were able to demonstrate that Giffen behavior exists for any set of preferences.
Using a combination of all of these techniques, John Davies published a paper (1994) discussing the same evidence first popularized by Paul Samuleson (1984) regarding Giffen behavior during the Irish potato famine. Davies specifies that the level of real income needed for Giffen behavior to occur is one that maintains biological subsistence, implying a very low level of income. Furthermore, he deduces that it was Ireland’s impoverished states that lead to binding subsistence constraints which in turn created a specialization in the consumption of potatoes, that he identified as Giffen behavior. Davies also establishes some of the difficulties behind gathering evidence of Giffen behavior and links this to the lack of available data on individual consumption. In addition to this he brings forth the idea that the extremely poor individuals that are most likely to display Giffenity are often found not only be consumers of “Giffen goods” but also their producers. This leads to a positive correlation between increasing prices and their income level rising making evidence of Giffenity among the poor hard to find. Walkers (1987) also made reference to this issue and used it to illustrate the complications regarding the empirical verification of Giffenity through data collection. Unfortunately, there have been economists that have proceeded to disagree with, or discredit Davies’ and Samuelson’s account of the potato famine. Read (2013) challenges their findings by suggesting that it was not the potato but the bacon pigs that demonstrated a Giffen type behavior among the wealthier consumers rather than the extremely poor. In addition to this, Sherwin Rosen (1999) and Dwyer and Lindsay (1984) altogether reject the example of the potato famine as evidence of Giffen behavior.
A second example of Giffen behavior emerging amid low income households is Jensen and Miller’s research paper (2008). Having previously conducted a since discredited study (2002), they published a second paper (2008) addressing the issues identified with their earlier report. Argued to be the first to have found empirical evidence of Giffen behavior (Zhu D., 2016), they provide compelling proof of Giffenity among the extremely poor in the Chinese provinces of Hunan and Gansu. They illustrate this using dietary staples such as rice, wheat and noodles and, among others, the following two circumstances: “households are poor enough that they face subsistence nutrition concerns” but “cannot be so impoverished that they consume only the staple good” (Jensen and Miller, 2008). In conclusion, they find that only low income consumers, strongly reliant on a staple good and with restricted substitution options, will display a positive price elasticity and as a result, exhibit Giffen behavior.
A final example observing positively sloped demand curves among poor households, was Battalio, Kagel and Kogut’s experimental confirmation of Giffen behavior using rats (1991). They repudiate, with the help of evidence, that Giffenity surfaces for high income individuals. However, this case has been subject to controversy regarding the use of animals to demonstrate human behavior.
On the other hand, Doi, Iwasa and Shimomura (2009) who adopt a utility function based approach, disagree that Giffenity exclusively occurs at low income levels. Using well-behaved utility functions defined as convex indifference curves enclosed in a commodity space, they differentiate them with respect to price and income to find that Giffen behavior arises when an arbitrarily low proportion of income is used on an inferior good. Therefore, they conclude that it is incorrect to assume that Giffenity is a “phenomenon that occurs only under extreme circumstances”, in other words, low income households. In addition, Haagsma (2012) uses a similar method, based on Wold (1948) and Slutsky (1956), to suggest that Giffen behavior is not restricted to low income households but exists at high income levels instead. He adopts symbolic notation as opposed to numerical examples to illustrate what preferences generate Giffen behavior. Contradicting the idea that Giffenity exists only under extreme poverty, he proposes that harmful activities such as smoking and drinking, where the desire for the good increases with its continued consumption, give rise to Giffenity regardless of income levels. Haagsma also criticizes the standard approach using subsistence examples to explain Giffen behavior, highlighting that “economic theory does not tell us that the specific properties of the indifference map that give rise to Giffen behavior only prevail at relatively low levels of utility” (Lipsey R. and Rosenbluth G., 1971; Dougan W., 1982). Lastly, a study examining whether one of Mexico’s staple foods, the tortilla, exhibited Giffen behavior, concluded that it was not a Giffen but an inferior good (Mckenzie D., 2001). This research goes against the idea that Giffenity arises for dietary staples, a concept developed by many of whom believe it to be the reason why Giffen behavior is a phenomenon of low income.
There remains another category of economists such as Nachabar (1998), McDonough and Eisenhauer (1995) and to some extent Koenker (1977) and Stigler (1947) who deny the very existence of Giffen behavior. They argue that the examples put forth do not present enough indisputable evidence to confirm the paradox’s existence.
There has been much controversy not only surrounding the verification of Giffen behavior but also whether it is a phenomenon of low income, high income or both. However, I believe that, under albeit uncommon conditions, it is a phenomenon of low income. All of the real life examples discussed above use a set of conditions containing either a poverty constraint or one that is closely linked, such as a nutritional, biological or rationing constraint. (Boland L., 1977; Gilley O. and Karels G., 1991; Jensen R. and Miller N., 2008)